linux-kernel.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
* [PATCH 3/6] gfs2: simplify current_tail() via list_last_entry_or_null()
@ 2013-11-15  4:32 Jeff Liu
  2013-11-15 16:39 ` Oleg Nesterov
  0 siblings, 1 reply; 4+ messages in thread
From: Jeff Liu @ 2013-11-15  4:32 UTC (permalink / raw)
  To: linux-kernel
  Cc: akpm, xfs, cluster-devel, linux-mtd, jfs-discussion, oleg, jiri, gregkh

From: Jie Liu <jeff.liu@oracle.com>

Simplify the code in current_tail() via list_last_entry_or_null().

Signed-off-by: Jie Liu <jeff.liu@oracle.com>
---
 fs/gfs2/log.c | 10 +++-------
 1 file changed, 3 insertions(+), 7 deletions(-)

diff --git a/fs/gfs2/log.c b/fs/gfs2/log.c
index 610613f..555f767 100644
--- a/fs/gfs2/log.c
+++ b/fs/gfs2/log.c
@@ -441,13 +441,9 @@ static unsigned int current_tail(struct gfs2_sbd *sdp)
 
 	spin_lock(&sdp->sd_ail_lock);
 
-	if (list_empty(&sdp->sd_ail1_list)) {
-		tail = sdp->sd_log_head;
-	} else {
-		tr = list_entry(sdp->sd_ail1_list.prev, struct gfs2_trans,
-				tr_list);
-		tail = tr->tr_first;
-	}
+	tr = list_last_entry_or_null(&sdp->sd_ail1_list, struct gfs2_trans,
+				     tr_list);
+	tail = tr ? tr->tr_first : sdp->sd_log_head;
 
 	spin_unlock(&sdp->sd_ail_lock);
 
-- 
1.8.3.2

^ permalink raw reply related	[flat|nested] 4+ messages in thread

* Re: [PATCH 3/6] gfs2: simplify current_tail() via list_last_entry_or_null()
  2013-11-15  4:32 [PATCH 3/6] gfs2: simplify current_tail() via list_last_entry_or_null() Jeff Liu
@ 2013-11-15 16:39 ` Oleg Nesterov
  2013-11-15 17:12   ` introduce list_get_first/last (Was: [PATCH 3/6] gfs2: simplify current_tail() via list_last_entry_or_null()) Oleg Nesterov
  0 siblings, 1 reply; 4+ messages in thread
From: Oleg Nesterov @ 2013-11-15 16:39 UTC (permalink / raw)
  To: Jeff Liu
  Cc: linux-kernel, akpm, xfs, cluster-devel, linux-mtd,
	jfs-discussion, jiri, gregkh

On 11/15, Jeff Liu wrote:
>
> @@ -441,13 +441,9 @@ static unsigned int current_tail(struct gfs2_sbd *sdp)
>  
>  	spin_lock(&sdp->sd_ail_lock);
>  
> -	if (list_empty(&sdp->sd_ail1_list)) {
> -		tail = sdp->sd_log_head;
> -	} else {
> -		tr = list_entry(sdp->sd_ail1_list.prev, struct gfs2_trans,
> -				tr_list);
> -		tail = tr->tr_first;
> -	}
> +	tr = list_last_entry_or_null(&sdp->sd_ail1_list, struct gfs2_trans,
> +				     tr_list);
> +	tail = tr ? tr->tr_first : sdp->sd_log_head;
>  

Personally I agree with Steven. At least in this case
list_last_entry_or_null() doesn't really help to simplify the code.

But probably list_last_entry() makes sense in the "else" branch,
athough this is minor.


Off-topic. Not sure this really makes sense, but I was thinking about

	list_get_first(pos, head, member)	\
		((pos) = list_first_entry(head, typeof(*pos), member))

and list_get_first() last of course. The obvious advantage is that
compared to

	tr = list_last_entry(sdp->sd_ail1_list, struct gfs2_trans, tr_list);

above you do not need to type "struct gfs2_trans",

	list_get_last(tr, sdp->sd_ail1_list, tr_list);

looks a bit better.

Oleg.


^ permalink raw reply	[flat|nested] 4+ messages in thread

* introduce list_get_first/last (Was: [PATCH 3/6] gfs2: simplify current_tail() via list_last_entry_or_null())
  2013-11-15 16:39 ` Oleg Nesterov
@ 2013-11-15 17:12   ` Oleg Nesterov
  2013-11-18  5:41     ` Jeff Liu
  0 siblings, 1 reply; 4+ messages in thread
From: Oleg Nesterov @ 2013-11-15 17:12 UTC (permalink / raw)
  To: Jeff Liu
  Cc: linux-kernel, akpm, xfs, cluster-devel, linux-mtd,
	jfs-discussion, jiri, gregkh

On 11/15, Oleg Nesterov wrote:
>
> But probably list_last_entry() makes sense in the "else" branch,
> athough this is minor.
>
>
> Off-topic. Not sure this really makes sense, but I was thinking about
>
> 	list_get_first(pos, head, member)	\
> 		((pos) = list_first_entry(head, typeof(*pos), member))
>
> and list_get_first() last of course. The obvious advantage is that
> compared to
>
> 	tr = list_last_entry(sdp->sd_ail1_list, struct gfs2_trans, tr_list);
>
> above you do not need to type "struct gfs2_trans",
>
> 	list_get_last(tr, sdp->sd_ail1_list, tr_list);
>
> looks a bit better.

And list.h can use it too.

So how about the patch below? I can't decide whether it cleanups or
uglfies the code, I do not trust my taste ;) But to me it looks a
bit better this way.

Oleg.
--

diff --git a/include/linux/list.h b/include/linux/list.h
index ef95941..f52aba8 100644
--- a/include/linux/list.h
+++ b/include/linux/list.h
@@ -372,6 +372,12 @@ static inline void list_splice_tail_init(struct list_head *list,
 #define list_last_entry(ptr, type, member) \
 	list_entry((ptr)->prev, type, member)
 
+#define list_get_first(pos, head, member)	\
+	((pos) = list_first_entry(head, typeof(*(pos)), member))
+
+#define list_get_last(pos, head, member)	\
+	((pos) = list_last_entry(head, typeof(*(pos)), member))
+
 /**
  * list_first_entry_or_null - get the first element from a list
  * @ptr:	the list head to take the element from.
@@ -443,7 +449,7 @@ static inline void list_splice_tail_init(struct list_head *list,
  * @member:	the name of the list_struct within the struct.
  */
 #define list_for_each_entry(pos, head, member)				\
-	for (pos = list_first_entry(head, typeof(*pos), member);	\
+	for (list_get_first(pos, head, member);				\
 	     &pos->member != (head);					\
 	     pos = list_next_entry(pos, member))
 
@@ -454,7 +460,7 @@ static inline void list_splice_tail_init(struct list_head *list,
  * @member:	the name of the list_struct within the struct.
  */
 #define list_for_each_entry_reverse(pos, head, member)			\
-	for (pos = list_last_entry(head, typeof(*pos), member);		\
+	for (list_get_last(pos, head, member);				\
 	     &pos->member != (head); 					\
 	     pos = list_prev_entry(pos, member))
 
@@ -517,7 +523,7 @@ static inline void list_splice_tail_init(struct list_head *list,
  * @member:	the name of the list_struct within the struct.
  */
 #define list_for_each_entry_safe(pos, n, head, member)			\
-	for (pos = list_first_entry(head, typeof(*pos), member),	\
+	for (list_get_first(pos, head, member),				\
 		n = list_next_entry(pos, member);			\
 	     &pos->member != (head); 					\
 	     pos = n, n = list_next_entry(n, member))
@@ -564,7 +570,7 @@ static inline void list_splice_tail_init(struct list_head *list,
  * of list entry.
  */
 #define list_for_each_entry_safe_reverse(pos, n, head, member)		\
-	for (pos = list_last_entry(head, typeof(*pos), member),		\
+	for (list_get_last(pos, head, member),				\
 		n = list_prev_entry(pos, member);			\
 	     &pos->member != (head); 					\
 	     pos = n, n = list_prev_entry(n, member))


^ permalink raw reply related	[flat|nested] 4+ messages in thread

* Re: introduce list_get_first/last (Was: [PATCH 3/6] gfs2: simplify current_tail() via list_last_entry_or_null())
  2013-11-15 17:12   ` introduce list_get_first/last (Was: [PATCH 3/6] gfs2: simplify current_tail() via list_last_entry_or_null()) Oleg Nesterov
@ 2013-11-18  5:41     ` Jeff Liu
  0 siblings, 0 replies; 4+ messages in thread
From: Jeff Liu @ 2013-11-18  5:41 UTC (permalink / raw)
  To: Oleg Nesterov
  Cc: jfs-discussion, jiri, gregkh, linux-kernel, xfs, cluster-devel,
	linux-mtd, akpm

Hi Oleg,

Sorry for my late response!

On 11/16/2013 01:12 AM, Oleg Nesterov wrote:

> On 11/15, Oleg Nesterov wrote:
>>
>> But probably list_last_entry() makes sense in the "else" branch,
>> athough this is minor.
>>
>>
>> Off-topic. Not sure this really makes sense, but I was thinking about
>>
>> 	list_get_first(pos, head, member)	\
>> 		((pos) = list_first_entry(head, typeof(*pos), member))
>>
>> and list_get_first() last of course. The obvious advantage is that
>> compared to
>>
>> 	tr = list_last_entry(sdp->sd_ail1_list, struct gfs2_trans, tr_list);
>>
>> above you do not need to type "struct gfs2_trans",
>>
>> 	list_get_last(tr, sdp->sd_ail1_list, tr_list);
>>
>> looks a bit better.
> 
> And list.h can use it too.
> 
> So how about the patch below? I can't decide whether it cleanups or
> uglfies the code, I do not trust my taste ;) But to me it looks a
> bit better this way.

I personally think this is better for GFS2 case if I understood correctly,
and it can make the existing list helper a bit neat per my taste :)

The current list_first_entry() works fine consider a common pattern below:
while (!list_empty(ptr)) {
	struct dummy *p = list_first_entry(ptr, struct dummy, member);

	....
}

But I'm not sure if we can find out more external use cases can get
benefits with list_get_first/last...

Thanks,
-Jeff

> 
> Oleg.
> --
> 
> diff --git a/include/linux/list.h b/include/linux/list.h
> index ef95941..f52aba8 100644
> --- a/include/linux/list.h
> +++ b/include/linux/list.h
> @@ -372,6 +372,12 @@ static inline void list_splice_tail_init(struct list_head *list,
>  #define list_last_entry(ptr, type, member) \
>  	list_entry((ptr)->prev, type, member)
>  
> +#define list_get_first(pos, head, member)	\
> +	((pos) = list_first_entry(head, typeof(*(pos)), member))
> +
> +#define list_get_last(pos, head, member)	\
> +	((pos) = list_last_entry(head, typeof(*(pos)), member))
> +
>  /**
>   * list_first_entry_or_null - get the first element from a list
>   * @ptr:	the list head to take the element from.
> @@ -443,7 +449,7 @@ static inline void list_splice_tail_init(struct list_head *list,
>   * @member:	the name of the list_struct within the struct.
>   */
>  #define list_for_each_entry(pos, head, member)				\
> -	for (pos = list_first_entry(head, typeof(*pos), member);	\
> +	for (list_get_first(pos, head, member);				\
>  	     &pos->member != (head);					\
>  	     pos = list_next_entry(pos, member))
>  
> @@ -454,7 +460,7 @@ static inline void list_splice_tail_init(struct list_head *list,
>   * @member:	the name of the list_struct within the struct.
>   */
>  #define list_for_each_entry_reverse(pos, head, member)			\
> -	for (pos = list_last_entry(head, typeof(*pos), member);		\
> +	for (list_get_last(pos, head, member);				\
>  	     &pos->member != (head); 					\
>  	     pos = list_prev_entry(pos, member))
>  
> @@ -517,7 +523,7 @@ static inline void list_splice_tail_init(struct list_head *list,
>   * @member:	the name of the list_struct within the struct.
>   */
>  #define list_for_each_entry_safe(pos, n, head, member)			\
> -	for (pos = list_first_entry(head, typeof(*pos), member),	\
> +	for (list_get_first(pos, head, member),				\
>  		n = list_next_entry(pos, member);			\
>  	     &pos->member != (head); 					\
>  	     pos = n, n = list_next_entry(n, member))
> @@ -564,7 +570,7 @@ static inline void list_splice_tail_init(struct list_head *list,
>   * of list entry.
>   */
>  #define list_for_each_entry_safe_reverse(pos, n, head, member)		\
> -	for (pos = list_last_entry(head, typeof(*pos), member),		\
> +	for (list_get_last(pos, head, member),				\
>  		n = list_prev_entry(pos, member);			\
>  	     &pos->member != (head); 					\
>  	     pos = n, n = list_prev_entry(n, member))
> 
> _______________________________________________
> xfs mailing list
> xfs@oss.sgi.com
> http://oss.sgi.com/mailman/listinfo/xfs



^ permalink raw reply	[flat|nested] 4+ messages in thread

end of thread, other threads:[~2013-11-18  5:41 UTC | newest]

Thread overview: 4+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2013-11-15  4:32 [PATCH 3/6] gfs2: simplify current_tail() via list_last_entry_or_null() Jeff Liu
2013-11-15 16:39 ` Oleg Nesterov
2013-11-15 17:12   ` introduce list_get_first/last (Was: [PATCH 3/6] gfs2: simplify current_tail() via list_last_entry_or_null()) Oleg Nesterov
2013-11-18  5:41     ` Jeff Liu

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).